OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

uddi-spec message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: RE: [uddi-spec] Summary of standards for external taxonomies


Matthew,

I do not recall now whether I had responded to your previous suggestion,
but from the outset I believe we shared largely complementary, if not
similar, visions of approaching this issue.  I thought of the
possibility you advocate, but decided in favor of the alternative I
ended up proposing, because I thought it would be simpler for others to
understand.  I completely agree with your proposal and perceive it as a
viable modeling possibility, which is more versatile and superior in
many respects to my own proposal.  The only problem might arise in
mapping the information provided by the external taxonomy service to
keyValues in UDDI keyedReferences (i.e. applying external service's keys
in UDDI), which is not specifically addressed in your solution.  This
should not be an issue for most 'normal' taxonomies, which have one
distinct identifier for each category.

Regards,
Daniel


> -----Original Message-----
> From: Matthew Dovey [mailto:matthew.dovey@las.ox.ac.uk] 
> Sent: Monday, March 03, 2003 6:01 PM
> To: Daniel Feygin; Max Voskob; uddi-spec
> Subject: RE: [uddi-spec] Summary of standards for external taxonomies
> 
> 
> Daniel,
> 
> I suggested something similar earlier in the month, but 
> different in a fundamental way that I think makes it more generic.
> 
> If I read you suggestion correctly you assume that TopicMap 
> taxonomies are a particular category of taxonomy. I don't 
> believe this is correct. A TopicMap is a particular 
> representation of a taxonomy but by no means the unique 
> representation.
> 
> e.g. a particular taxonomy (such as the industry 
> classifications already in UDDI) could be represented as a 
> TopicMap, in a DAML/OIL schema, in Max's schema etc.
> 
> My suggestion was that we introduce an tModel to represent an 
> external Taxonomy service (I'll call this 
> tModel_ExternalTaxonomy). This external taxonomy service for 
> a given category (I'll say this is represented by
> tModel_OurCategory) is then registered in UDDI as any other 
> WebService. However, its bindingTemplate would have a tModel 
> fingerprint of
> 
> tModel_ExternalTaxonomy, tModel_OurCategory plus tModels to 
> represent how to access the External Taxonomy.
> 
> At this stage we would introduce tModels indicating httlp 
> access to TopicMaps, DAML/OIL etc. I don't know whether this 
> should be a single tModel (e.g. tModalTopicMapviaHTTP, 
> tModelDAMLviaHTTP) or a pair one to indicate the schema the 
> other the access mechanism (e.g. tModelTopicMap, tModelHTTP 
> or tModelDAML, tModelFTP).
> 
> This mechanisms allows us to easily drop in tModels for 
> richer API's if/when such become available. It also means 
> that you may find (and use) different taxonomic 
> services/representations for the same category.
> 
> Matthew
> 
> 
> 
> 
> -----Original Message-----
> From:	Daniel Feygin [mailto:feygin@unitspace.com]
> Sent:	Mon 03/03/2003 09:18
> To:	'Max Voskob'; 'uddi-spec'
> Cc:	
> Subject:	RE: [uddi-spec] Summary of standards for external
> taxonomies
> 
> Max, TC,
>  
> With regard to taxonomy representations, I suggest the 
> following modeling approach (I will use Topic Maps for 
> illustration).  First, a categorization tModel is created to 
> represent all Topic Maps taxonomies. Specific topics/topic 
> maps are then represented as separate tModels (lets call them 
> Topic Maps instances) categorized by the Topic Maps 
> categorization tModel.  One of the overviewDocs of those 
> Topic Maps instances should point to the XTM or ISO/IEC 13250 
> document describing the taxonomy structure ("associations" in 
> the case of Topic Maps).
>  
> These Topic Maps instance tModels could then be used in 
> categoryBags to categorize information in the registry with 
> their keyReferences' keyValues containing the value of the 
> topic's ID.  If it is impossible to identify a source 
> attribute to assign to the keyValue attribute, then the Topic 
> Maps categorization tModel should provide documentation on 
> the coding method for keyValue based on attributes of Topic 
> Maps representation schema.  This method should reliably 
> generate unique values within the key space of each Topic 
> Maps instance.
>  
> This assessment is based purely on the information provided 
> by Max (and my assumptions about the untold) and may be 
> unimplementable due to obstacles beyond Max's review.
>  
> Generalizing this approach for all other taxonomy 
> representation formats, we can support a limited subset of 
> taxonomy functions (no API's
> yet) identified at the FTF by delegating them to taxonomy 
> specialists and simultaneously providing a way to integrate 
> their work into UDDI.
>  
> Daniel
>  
> 
> -----Original Message-----
> From: Max Voskob [mailto:mvoskob@msi.com.au] 
> Sent: Saturday, March 01, 2003 11:54 AM
> To: uddi-spec
> Subject: [uddi-spec] Summary of standards for external taxonomies
> 
> 
> Hi all ;)
>  
> Me (Max) and Matthew prepared a short summary about possible 
> standards for external taxonomies. 
> Please, find a minute and familiarise yourself with what's 
> available at the moment. 
> If you think that any of the proposed standards is the way to 
> go - speak up. If you are aware about another standard that 
> is not included in the doco
> - speak up too. 
>  
> Cheers,
> Max
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 


----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>




[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]